Examining L2 speakers’ expression of opinion in the Trinity Lancaster Corpus

Jung, Yejin and Brezina, Vaclav and Gablasova, Dana (2024) Examining L2 speakers’ expression of opinion in the Trinity Lancaster Corpus. PhD thesis, Lancaster University.

Text (2024jungphd)
2024jungphd.pdf - Published Version
Restricted to Repository staff only until 3 February 2030.
Available under License Creative Commons Attribution-NonCommercial.
Download (3MB)

Text (2024jungphdinternal)
2024jungphdinternal.pdf - Published Version
Restricted to Repository staff only until 3 February 2030.
Available under License Creative Commons Attribution-NonCommercial.
Download (3MB)

Abstract

This PhD thesis is motivated by the following research needs identified in the context of Second Language Acquisition and Second Language Teaching and Testing. First, opinion-giving is one of the most-represented tasks in teaching and assessing L2 speaking (e.g., Brown & Ducasse, 2019) but a limited number of studies has provided empirical evidence on how L2 speakers express their opinion in interactive speech. Second, L2 speakers’ oral skills have been studied with respect to pragmatic and interactional features (e.g., responding to a previous speaker in Lam, 2018), which need further investigation on whether they would vary depending on learners’ backgrounds (e.g., L2 proficiency level or L1 background), and if so, how. In response to these research gaps, the current study aims to i) describe different types of opinion statements (OS, henceforth) produced by L2 English speakers in interactive speech and ii) evaluate the effect of L2 proficiency and L1 background on the occurrence of different types of OSs. The different types of OSs examined with respect to the research aims concern three aspects: i) whether an OS is simply stated without any supportive statements (e.g., giving a reason for a previously stated idea) (Simple OS) or both stated and supported (Complex OS); ii) whether an OS is requested by an interlocutor (Prompted OS) or occurs without an interlocutor’s direct request for opinion (Unprompted OS); and iii) what type of supportive statement is provided (e.g., justifying, elaborating). Taking a quantitative corpus-based approach, the study sampled 286 transcripts of learners’ spoken performance during the GESE, a high-stakes exam of L2 English. Each exam involves a dialogue between a test-taker (L2 speakers of English) and an examiner (L1 speaker of English) and the transcripts are part of the Trinity Lancaster Corpus (Gablasova et al, 2019). The L2 speakers included in the dataset were collected from three different proficiency levels in L2 - low-intermediate (equivalent to the B1 level in CEFR), high-intermediate (B2 level), and advanced (C) - as well as two different L1 backgrounds - Spanish and Chinese. In the corpus data, the different types of OSs mentioned above were identified and classified according to the annotation scheme that this study has developed, informed by Appraisal Theory (Martin & White, 2005). The study utilised descriptive statistics to describe the patterns in the occurrence of OSs, and inferential statistics (One-way ANOVA, Two-way ANOVA, and T-test) to identify the effect of the two learner variables on the occurrence of the features. Results showed that L2 speakers generally produced a higher number of Complex OSs than Simple OSs, and a higher number of Unprompted OSs than Prompted OSs. It was also found that they supported OSs predominantly by giving a reason for a previously-stated OS, elaborating on it, and giving a contrasting perspective to it. A significant effect of L2 proficiency level was identified on expression of opinion: the frequency of Complex OSs increased significantly from the intermediate levels (B1 and B2) to the advanced level (C); but that of Simple OSs decreased significantly. Both Prompted and Unprompted OSs increased from the intermediate levels (B1 and B2) to the advanced level (C). There was, however, little effect of L1 background on expression of opinion. The findings suggest a tendency for L2 speakers to express opinion in elaborate unsolicited manners in the spoken discourse examined in the current study. Although this tendency became stronger towards higher proficiency levels, it was also affected by interlocutors (examiners)’ utterances. The patterns are discussed from the perspective of L2 speakers’ pragmatic strategies to maintain politeness as well as considering the nature and discourse of the GESE exam. The findings contribute to a theoretical understanding of speaking competence of L2 speakers from various backgrounds as well as provide implications for the practice of teaching and assessing opinion-giving skills in co-constructive interactions.